|Home | About | Journals | Submit | Contact Us | Français|
Background.Limited data exist on the immunogenicity of the 2009 influenza A (H1N1) vaccine among immunocompromised persons, including those with human immunodeficiency virus (HIV) infection.
Methods.We compared the immunogenicity and tolerability of a single dose of the monovalent 2009 influenza A (H1N1) vaccine (strain A/California/7/2009H1N1) between HIV-infected and HIV-uninfected adults 18–50 years of age. The primary end point was an antibody titer of ≥1:40 at day 28 after vaccination in those with a prevaccination level of ≤1:10, as measured by hemagglutination-inhibition assay. Geometric mean titers, influenza-like illnesses, and tolerability were also evaluated.
Results.One hundred thirty-one participants were evaluated (65 HIV-infected and 66 HIV-uninfected patients), with a median age of 35 years (interquartile range, 27–42 years). HIV-infected persons had a median CD4 cell count of 581 cells/mm3 (interquartile range, 476–814 cells/mm3) , and 82% were receiving antiretroviral medications. At baseline, 35 patients (27%) had antibody titers of >1:10. HIV-infected patients (29 [56%] of 52), compared with HIV-uninfected persons (35 [80%] of 44), were significantly less likely to develop an antibody response (odds ratio, .20; P = .003). Changes in the median geometric mean titer from baseline to day 28 were also significantly lower in HIV-infected patients than in HIV-uninfected persons (75 vs 153; P = .001). Five influenza-like illnesses occurred (2 cases in HIV-infected persons), but none was attributable to the 2009 influenza H1N1 virus. The vaccine was well tolerated in both groups.
Conclusions.Despite high CD4 cell counts and receipt of antiretroviral medications, HIV-infected adults generated significantly poorer antibody responses, compared with HIV-uninfected persons. Future studies evaluating a 2-dose series or more-immunogenic influenza A (H1N1) vaccines among HIV-infected adults are needed (ClinicalTrials.gov NCT00996970).
Initial studies showed that a single dose of the novel inactivated 2009 influenza H1N1 vaccine was immunogenic in ~95% of healthy adults [ 1, 2], resulting in its global production and distribution. Additional studies have shown that a 2-dose series is immunogenic in healthy children and older adults (>60 years of age) [ 3, 4]. However, to date, few studies have been published on the immunogenicity of this vaccine in immunocompromised adults, including those with human immunodeficiency virus (HIV) infection.
HIV-infected persons are at high risk of influenza-related complications and death [ 5– 9]. The Centers for Disease Control and Prevention recommended administering the 2009 influenza H1N1 vaccine to adults at risk of influenza-related complications, including those with HIV infection [ 10]. Prior studies have shown that HIV-infected persons may generate lower immune responses to vaccinations, including to the seasonal influenza vaccine, compared with healthy adults [ 11– 14].
Because influenza is a major cause of respiratory illness in this population [ 15] and vaccination is the primary method of prevention, clinical data on the immunogenicity of the 2009 H1N1 vaccine in HIV-infected persons are needed. We conducted a prospective study of the immunogenicity and tolerability of an influenza A (H1N1) 2009 monovalent vaccine in HIV-infected adults, compared with HIV-uninfected adults.
We compared the immunogenicity of a monovalent 2009 influenza A (H1N1) vaccine (strain A/California/7/2009[H1N1]; Novartis Vaccines and Diagnostics) between HIV-infected adults and HIV-uninfected adults. The vaccine manufacturer was not involved in the study in any capacity.
Participants receiving the 2009 H1N1 vaccination were screened for eligibility and provided written informed consent. Both HIV-infected and HIV-uninfected groups were enrolled simultaneously from 29 October through 2 December 2009. The study was approved by a central military institutional review board, was conducted in accordance with the principles of the Declaration of Helsinki and standards of Good Clinical Practice (as defined by the International Conference on Harmonization), and was registered with the Clinical Trials network (NCT00996970).
HIV-infected participants had documented HIV infection (positive enzyme-linked immunosorbent assay [ELISA] and Western blot results) at enrollment. HIV-uninfected persons had a negative HIV ELISA result ≤1 year before enrollment. Inclusion criteria for all participants were age 18–50 years and no serious medical conditions, except for the diagnosis of HIV infection in the former group. Exclusion criteria included documented pregnancy or being ≤6 weeks postpartum, being a health care worker involved in direct patient care (because of higher risk of influenza H1N1 exposure), acute febrile illness ≤30 days before vaccination, diabetes mellitus, systemic steroid or immunosuppressive medication use ≤4 weeks before vaccination, cancer (except nonmelanoma skin cancer), history of organ transplantation, chronic active hepatitis B or C, current illicit drug use or alcohol abuse, blood transfusion within the previous year, allergy to eggs, significant adverse reaction to a prior vaccination, receipt of another vaccination within ≤4 weeks (seasonal influenza vaccination allowed), or history of confirmed or suspected 2009 influenza H1N1 infection.
The primary objective of the study was to compare rates of seroconversion, defined as a geometric mean titer (GMT) of ≥1:40 after vaccination (day 28) in persons with a prevaccination (day 0) level of ≤1:10, as measured by hemagglutination-inhibition assay (HAI) [ 2, 16, 17], between HIV-infected adults and HIV-uninfected adults. To examine antibody levels generated by naive persons without prior immunity to the 2009 influenza H1N1 strain, we excluded persons with a baseline GMT >1:10 for our primary objective. Secondary outcomes, which evaluated all study participants, included achieving seroprotection, defined as a postvaccination titer ≥1:40, a ≥4-fold increase in titers from before to after vaccination (regardless of postvaccination level achieved), change in the GMT, and clinical events, including influenza-like illness (ILI). In addition, the tolerability of the vaccine and its impact on HIV RNA levels and CD4 cell counts in HIV-infected persons were evaluated.
The study was designed on the basis of the primary objective and assumed that 90% HIV-uninfected persons and 60%–70% of HIV-infected persons would experience seroconversion [ 1, 2, 18– 20]. With use of a 2–independent proportions Fisher's exact test with power of 85% and 1-sided α=.05 (superiority alternative hypothesis: HIV-uninfected persons will have a higher proportion with seroconversion, compared with HIV-infected patients), a sample size of 100 participants was estimated. Because an estimated 32% of participants [ 2] might have been exposed to 2009 influenza H1N1 before vaccination and might have had pre-existing antibodies, the sample size was increased to 132 participants. Because vaccine responses may vary by age, we used frequency matching to assure comparability between the study arms, creating 4 groups (each with 33 participants) stratified by HIV status and age ≥35 years or <35 years (the midpoint between ages 18 and 50 years).
Vaccines were stored and administered in accordance with manufacturers’ guidelines. Participants completed self-administered questionnaires regarding medical history, medication use, self-reported height and weight for body mass index (BMI), current cigarette use, history of influenza, seasonal influenza vaccinations within the previous 3 years, and number of household members. Study coordinators used medical records to collect relevant demographic characteristics and medical history, including HIV infection history.
All participants were observed for ~30 min after vaccination, and immediate reactions were recorded. A diary card was provided to record solicited local and systemic adverse reactions and any unsolicited reactions within 7 days after vaccination. Adverse reactions were graded on the basis of impact on daily activities as mild (not interfering with usual function), moderate (interfering to some extent with usual function), and severe (significantly interfering with usual function) [ 21]. Reactions possibly related to vaccination that were life-threatening or that resulted in hospitalization or disability were reported to the Vaccine Adverse Event Reporting System.
Participants who experienced ILI events (defined as a temperature of ≥37.8 °C and cough and/or a sore throat in the absence of another known cause) [ 22] were asked to have nasopharyngeal swab samples obtained for rapid influenza antigen testing and real-time reverse-transcriptase polymerase chain reaction for the 2009 influenza H1N1 virus [ 23] and other select pathogens (adenovirus, rhinovirus, Mycoplasma pneumoniae, Streptococcus pneumoniae, Bordetella pertussis, Legionella pneumophilia, Chlamydophila pneumoniae, and respiratory syncytial virus). Information on exposures to persons with ILI during the study period was also collected.
Serum samples for influenza-specific antibody responses were collected at day 0 and day 28 (±4 days). CD4 cell counts (determined by flow cytometry) and plasma HIV RNA levels (Roche Amplicor; lower limit of detection, 50 copies/mL) were determined for HIV-infected participants at days 0 and 28. Antibody responses were detected using HAI with 2-fold serial dilutions, as described elsewhere [ 24, 25], at the Naval Medical Research Center (Silver Spring, MD). HAI assays were conducted using .5% turkey erythrocytes; the reference antiserum samples were supplied by the Centers for Disease Control and Prevention. Serum samples were treated with receptor-destroying enzyme, heme-adsorbed, and tested in duplicates in 2 independent assays, with the GMT reported as the final titer. For computational purposes, titers of <1:10 were assigned a value of 1:5, and those >1:1280 were assigned a value of 1:1280.
Descriptive statistics (median values, interquartile ranges [IQRs], counts, or proportions) and unadjusted group comparisons (2-sample Student’s t tests and Fisher's exact tests) are shown. For multivariate adjusted analyses of the primary and secondary objectives, logistic regression for binary outcomes and linear regression for continuous outcomes were used. Adjusted models included age, race, number of seasonal influenza vaccinations in previous 3 years, and number of household members. Similar analyses were performed post-hoc for participants <35 years of age, participants ≥35 years of age, and participants who received the inactivated 2009 seasonal vaccine.
Exploratory analyses examining predictors of H1N1 seroconversion in HIV-infected participants were performed using simple logistic or linear regression models, as appropriate. Predictors considered included demographic characteristics, BMI, number of household members, number of seasonal influenza vaccinations in the previous 3 years, self-reported prior influenza illness, HIV infection duration, baseline and nadir CD4 cell counts, HIV RNA level, and receipt of highly active antiretroviral therapy (HAART).
Reported P values for prespecified primary and secondary outcomes are 1-sided in accordance with study design, and all other P values are 2-sided; P <.05 was considered to indicate statistical significance. All analyses were conducted using SAS, version 9.1 (SAS Institute).
A total of 132 participants were enrolled (66 HIV-infected and 66 HIV-uninfected persons); 131 (99%) completed both study visits, with 1 HIV-infected patient failing to schedule the second visit. Participants had a median age of 35 years (IQR, 27–42 years), 91% were male, and race included white (60%), black (24%), and other (16%) ( Table 1). Ninety-five percent of participants had received a 2009 seasonal influenza vaccine previously or at enrollment; 85% of these participants received an inactivated formulation. Seventy percent had received seasonal influenza vaccinations during each of the previous 3 years. Fourteen percent self-reported a diagnosis of seasonal influenza during their lifetime. Among HIV-infected persons, the median duration of HIV infection was 6.6 years (IQR, 2.1–13.5 years), the median CD4 cell count was 581 cells/mm3 (IQR, 476–814 cells/mm3), 57% had an undetectable HIV RNA level, and 82% were receiving HAART at the time of vaccination.
HIV-infected and HIV-uninfected participants were similar with regard to demographic characteristics and influenza illness history ( Table 1). HIV-infected persons had fewer household members than did persons in the HIV-uninfected arm. Although similar percentages of participants received the 2009 seasonal vaccination, HIV-infected persons were less likely to have received a live attenuated formulation and received fewer seasonal influenza vaccines during the previous 3 years.
At baseline, 35 participants (27%) had antibody titers of >1:10 by HAI, with no statistically significant differences between the HIV-infected group (13 [20%]) and the HIV-uninfected group (22 [33%]; P = .11). The primary end point was achieved by significantly fewer HIV-infected patients (56%) than HIV-uninfected adults (80%; adjusted odds ratio [OR], .20; P = .003) ( Table 2). Vaccine responses in all participants, regardless of prevaccination titer, for (1) a ≥4 fold increase in titer from before to after vaccination and (2) for seroresponse, defined as a level of ≥1:40 at day 28 after vaccination, also revealed significantly poorer responses in HIV-infected persons than in HIV-uninfected persons ( Table 2). We also examined the increase in GMT from baseline to day 28; the change was significantly smaller in HIV-infected persons (median, 75; IQR, 3–155) than in HIV-uninfected persons (median, 153; IQR, 62–376; P = .001).
Antibody responses for achieving the primary end point were evaluated stratified by age and revealed that differences in HIV-infected persons, compared with HIV-uninfected participants, were mainly in those ≥35 years of age ( Table 2). We also compared HIV-infected persons with HIV-uninfected persons restricted to those having received an inactivated 2009 seasonal influenza vaccine and found similar results, although they were not statistically significant, likely because of the reduced sample size ( Table 2).
There were 5 cases of ILI during the 28-day period after vaccination. Three occurred in HIV-uninfected patients, and 2 occurred in HIV-infected subjects. All patients with ILI underwent testing for seasonal and 2009 H1N1 influenza virus infection (with the exception of a single HIV-infected enrollee), and all had negative results. Results of testing for other respiratory pathogens were also negative. The median duration of ILI symptoms was 7 days (range, 3–16 days), and ILI self-resolved without antimicrobial agents. Of the 5 patients with ILI, 2 (1 in each arm) had prevaccination antibody levels of >1:10. All patients with ILI had a 4-fold increase in antibody levels after vaccination, except for 1 HIV-uninfected person. The number of participants reporting exposure to persons with ILI from day 0 through day 28 was similar between HIV-infected and HIV-uninfected arms (18 and 22, respectively; P = .57).
Exploratory analyses of factors associated with antibody responses were performed for HIV-infected persons ( Table 3). HIV-infected participants who achieved the primary end point, compared with those who did not, were more likely to be younger (median age, 31 vs 45 years; OR, .93; P = .01) and had a shorter duration of HIV infection (median duration, 5.7 vs 13.4 years; OR, .9; P = .007). There was an association between self-reported prior influenza illness and achieving an antibody response that had borderline statistical significance (median percentage, 24% vs 4%; OR, 7.0; P = .06). Race, sex, BMI, seasonal influenza vaccine history, number of household members, CD4 cell counts, HIV RNA levels, and receipt of HAART were not significantly associated with vaccine response.Similar overall findings were found in an examination of secondary end points for antibody responses (data not shown).
The vaccine was generally well-tolerated in both HIV-infected and HIV-uninfected persons. No immediate vaccine reactions were noted. During the first 7 days after vaccination, 79 participants (60%) experienced at least 1 solicited adverse reaction ( Table 4). Most adverse reactions were mild; the most common reaction was pain at the vaccine injection site (33%), followed by headache (30%) and malaise (24%). Unsolicited events were uncommon and are shown in Table 4. The number, severity, and type of all adverse reactions were similar between HIV-uninfected and HIV-infected persons. The occurrence of an adverse reaction was similar in participants who experienced seroconversion and in those who did not experience seroconversion (P >.99).
Only 1 participant (in the HIV-uninfected arm) developed a serious adverse event, possibly related to the vaccine—angioedema beginning on day 1 after vaccination (the participant also received the seasonal influenza vaccine on the same day) that lasted for 17 days and subsequently resolved after intensified antihistamine therapy.
No statistically significant changes in the plasma HIV RNA levels or CD4 cell counts were noted after vaccination. At day 28, the median change in the plasma log10 HIV RNA level was .0 copies/mL (IQR, −.1–0 copies/mL; P = .15), which is well within the anticipated variability of the test (.3 log). There was also no statistically significant difference in the proportions of participants with a detectable viral load between the visits (P = .17). Finally, no statistically significant change in the median CD4 cell count between visits (18 cells/mm3; IQR, −89–95 cells/mm3) was noted (P = .63).
We revealed that HIV-infected adults are significantly less likely to generate antibody responses to a single dose of the monovalent 2009 influenza A (H1N1) vaccine than are HIV-uninfected adults. Only 56% of HIV-infected persons experienced seroconversion, compared with 80% of HIV-uninfected persons. This finding suggests that HIV-infected adults may require a more immunogenic H1N1 vaccine or additional doses to achieve response levels seen in the general population.
Influenza virus strains are a major cause of respiratory illness in the 33 million persons infected with HIV worldwide [ 5, 7, 15]. With the rapid emergence of the 2009 influenza H1N1 pandemic, there were concerns that this novel virus may significantly affect this population; thus, the effectiveness of vaccines developed against novel influenza strains for HIV-infected persons is of clinical interest.
This study is one of the first to examine the immunogenicity of the 2009 influenza A (H1N1) vaccine in HIV-infected persons. Despite having robust CD4 cell counts (median, 581 cells/mm3), high HAART coverage, and few comorbid diseases, HIV-infected participants generated immune responses to vaccination at a lower frequency and magnitude than did HIV-uninfected persons. Prior studies have shown that HIV-infected patients often generate poorer responses to vaccination, compared with healthy adults [ 11– 14]. For example, seroconversion rates for seasonal influenza vaccination among HIV-infected persons were 30%–40%, albeit with a wide range of 27%–78%, depending on the study population [ 11– 14, 18, 19, 26– 34]. Potential reasons for poorer vaccine responses despite preserved CD4 cell counts include residual immune dysregulation affecting both T cell and B cell quantities and function, immune activation, and/or immunosenescence [ 35– 37].
Although immune responses generated are often lower in HIV-infected persons, it is unknown whether the influenza H1N1 vaccine may be sufficiently protective against clinical disease. For example, studies on seasonal influenza vaccination have shown modest protective efficacy despite lower than normal antibody levels [ 11– 13, 15], suggesting potential benefit to vaccination despite antibody levels achieved and advocating for its use in clinical practice. In our study, there were no cases of 2009 H1N1 influenza after vaccination; however, the study was not designed to evaluate efficacy, the follow-up period was only 28 days, and influenza H1N1 activity in the community dramatically decreased during the study period [ 38]. Although evaluation of all ILI events during the 6-month period after vaccination is in progress, further studies are needed to more fully elucidate the protective efficacy of this new vaccine among HIV-infected persons.
In HIV-infected persons in our study, predictors for antibody responses included age and duration of HIV infection; only 36% of HIV-infected participants ≥35 years of age achieved the primary end point, compared with 74% of those aged <35 years. Other studies evaluating seasonal influenza vaccines and 2009 influenza A (H1N1) vaccines in HIV-uninfected adults found age to be predictive of antibody responses [ 3, 39]. Our association between longer duration of HIV infection and poorer vaccine responses is unique, potentially related to increasing age and/or long-term immunosuppression. Prior studies, but not ours, have found that higher CD4 cell counts, lower plasma HIV RNA levels, and HAART use may be associated with improved responses to seasonal influenza vaccination [ 6, 19, 33, 34]. Our study may be limited by small numbers of late-stage patients and high HAART coverage.
We noted a novel relationship between lifetime history of an influenza illness and improved responses to influenza H1N1 vaccination, which may represent cross-priming from preserved immune responses to conserved components of the virus. Because this association is novel; influenza history was self-reported and, thus, limited by recall bias; and the ORs had large confidence intervals, confirmation by other studies is needed. We did not find statistically significant associations between receipt of seasonal influenza vaccinations during the previous 3 years and vaccine responses, similar to other studies [ 40, 41], although most participants had received vaccination annually.
The 2009 H1N1 influenza vaccine was well tolerated among participants. Although 60% of participants experienced ≥1 solicited adverse reaction, most reactions were mild. Numbers and types of reactions in our study are comparable to those in other studies examining the 2009 H1N1 and seasonal influenza vaccinations [ 1, 2, 42, 43]. There were no statistically significant differences in the number or severity of reactions between HIV-infected and HIV-uninfected persons. Furthermore, HIV-infected persons did not experience significant changes in plasma HIV RNA levels or CD4 cell counts, but our study may have had limited ability to detect such changes because of timing of measurements and the suppressed viral status of most participants. Theoretical concerns over potential immune cell activation have been cited with vaccination, but any changes have been transient [ 44– 49].
Our study had potential limitations. We evaluated a well-controlled cohort of HIV-infected persons and, thus, could not determine the impact of severe immunosuppression (eg, CD4 cell counts <200 cells/mm3) on H1N1 vaccine responses. However, our study had the advantage of evaluating early-diagnosed and -treated HIV-infected persons and concurrently evaluating a group of HIV-uninfected persons with similar baseline characteristics. Second, 27% of participants had a baseline titer of >1:10 (similar to other studies) [ 2] and could not be evaluated for our primary objective; however, our sample size was calculated to assure adequate power, and we evaluated several secondary objectives to examine all participants. Third, although we used standard HAI methodology [ 24, 25] to define seroresponses, similar to other studies evaluating 2009 H1N1 and seasonal influenza vaccinations [ 1, 2, 11, 16, 17], the protective level for immunity from influenza illness is unknown.
In summary, HIV-infected adults tolerated the monovalent 2009 influenza A (H1N1) vaccine well but generated significantly poorer antibody responses compared with HIV-uninfected adults. Because HIV-infected patients are experiencing longer life expectancies and because high CD4 cell counts and antiretroviral therapy do not appear to negate the impact of HIV infection on vaccine responses, additional studies evaluating alternate vaccine preparations or doses in HIV-infected persons are needed.
We thank Brian Agan, MD; Mary Bavaro, MD; Braden Hale, MD; Michelle Linfesty; Connor Eggleston; Barbara Nagaraj; Sara Echols, RN; Sheila Medina, MPH; Erin McDonough; Jean Vita; Heather Hairston; Anna Mason; and Gabriela Sanguineti, for assisting with study conduct; and Maryna Eichelberger, PhD, and Kathy Hancock, PhD, for assistance with HAI assay development. The content of this publication is the sole responsibility of the authors and does not necessarily reflect the views or policies of the National Institutes of Health, the Department of Health and Human Services, the Department of Defense, or the Departments of the Army, Navy, or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the US Government.
Financial support.This work was supported by the Infectious Disease Clinical Research Program, a Department of Defense program executed through the Uniformed Services University of the Health Sciences; the National Institute of Allergy and Infectious Diseases, National Institutes of Health (under Interagency Agreement Y1-AI-5072); and the Armed Forces Health Surveillance Center's Global Emerging Infections System (via project I204_10).
Potential conflicts of interest.All authors: no conflicts.